United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-454/R-92-021
October 1992
AIR
& EPA
A MODELING PROTOCOL FOR APPLYING MESOPUFF H
TO LONG RANGE TRANSPORT PROBLEMS
-------
EPA-454/R-92-021
A MODELING PROTOCOL FOR APPLYING MESOPUFF II
TO LONG RANGE TRANSPORT PROBLEMS
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
October 1992
-------
DISCLAIMER
This report has been reviewed by the Office of Air Quality
Planning and Standards, EPA, and approved for publication.
Mention of trade names or commercial products is not
intended to constitute endorsement or recommendation for
use.
-------
CONTENTS
Section Page
Figures ii
Tables iii
Preface iv
1.0 Introduction 1
2.0 Background 5
2.1 Role of Long Range Transport Models 5
2.2 Description of MESOPUFF H 8
3.0 Recommended Procedures for Applying MESOPUFF II 13
3.1 Spatial and Temporal Scales of Analysis 14
3.1.1 Spatial Scale 14
3.1.2 Temporal Scale 17
3.2 Compilation of Meteorological Data Bases 19
3.3 Application of MESOPUFF II Preprocessors 21
3.3.1 Application of READ56 22
3.3.2 Application of MESOPAC H 23
3.4 Application of MESOPUFF II 29
3.5 Control Strategy Evaluation 39
4.0 Example MESOPUFF II Application 43
4.1 Description of Example Problem 43
4.2 Preprocessor Applications 45
4.3 Application of MESOPUFF II 52
4.4 Summary of Results 56
5.0 References 63
Appendix A Example Input Data Sets for READ56, MESOPAC II
and MESOPUFF II A-l
-------
FIGURES
Number Page
2-1 Schematic Representation of Puff Superposition Model 9
3-1 Example Grid Configurations 16
4-1 Modeling Region and Source Location 44
4-2 Locations of Surface and Upper Air Stations 47
4-3 Schematic Representation of Receptor Networks 54
11
-------
TABLES
Number Page
3-1 Summary of MESOPAC H Run Control Inputs 25
3-2 Land Use Categories Used in MESOPAC H 26
3-3 Summary of Default Procedures Recommended
for Regulatory Applications of MESOPAC H 28
3-4 Summary of MESOPUFF H Run Control Inputs 33
3-5 Summary of Default Procedures Recommended
for Regulatory Applications of MESOPUFF H 38
4-1 Source Data for Example Problem 46
4-2 List of Surface and Upper Air Stations 49
4-3 Primary and Alternate Upper Air Stations 50
4-4 Summary of Computer Resources Used
for Monthly Simulations 55
4-5 Greatest Regionwide Impacts 57
4-6 Top Ten High/Second-high Predicted Concentrations:
Regionwide Assessment 58
4-7 Summary of Model Predictions for PSD Assessment 60
in
-------
PREFACE
This report summarizes procedures for applying
MESOPUFF II to regulatory problems dealing with the long
range transport of relatively inert pollutants. These
procedures were developed using a main frame version of the
model dated 85360. While newer versions of the model have
been constructed, the procedures and recommendations
discussed within this document should still generally be
applicable.
IV
-------
1.0 INTRODUCTION
Several environmental problems may occur as a result of the transport of air
pollutants over long distances. For example, air emissions from a source located in one
political jurisdiction may be impeding progress towards attaining the National Ambient
Air Quality Standards (NAAQS) in a different political jurisdiction. Similarly, a
proposed new source may cause deterioration in air quality at a remote, pristine area far
removed from the source. From a regulatory perspective, these types of problems must
be addressed by first quantifying the air quality impacts of existing or proposed sources
and then determining the appropriate level of emission control that is needed to mitigate
those impacts. One approach for performing these types of assessments entails applying
an air quality dispersion model to determine the source-receptor relationships associated
with long range transport. This document presents a protocol for applying one such
model, the MESOPUFF II dispersion model, to long range transport problems within a
regulatory framework.1'2
Straight-line Gaussian air quality dispersion models have traditionally been used to
identify specific source-receptor relationships for transport distances up to about 50km.
These models are not appropriate for assessing air quality impacts over longer distances
however. They do not adequately simulate long range plume transport and dispersion
primarily because they do not account for temporal variations in plume transport direction
nor vertical separation of pollutant plumes caused by diurnal changes in the depth of the
mixed layer. MESOPUFF II, for which this protocol has been developed, has been
designed to address these phenomena. It was selected because it meets several criteria
-------
for refined modeling techniques that are outlined in the "Guideline on Air Quality
Models".* Specifically,
1) the model is computerized and documented in a user's guide that
identifies the mathematical algorithms used in the model, the data
requirements of the model, and the program operating characteristics;
2) the model is accompanied with a complete data set for testing; and
3) model performance evaluations have been conducted with the model
that compare model predictions with observations.4
The protocol presented in this document describes recommended procedures for
applying MESOPUFF II to regulatory problems dealing with the long range transport of
relatively inert pollutants such as sulfur dioxide (SO2) and particulate matter. Procedures
are proposed for developing the data bases necessary to apply the models, formulating
the required model inputs, selecting appropriate model options for the application, and
applying the model and processing the output produced by the model. These
recommended procedures have evolved principally through experience gained in applying
the model, and from results obtained by conducting performance evaluations and
sensitivity tests. In addition, the protocol follows general modeling recommendations that
are contained in the "Guideline on Air Quality Models"3 wherever applicable.
As noted above, the protocol is applicable to relatively inert pollutants, and as such,
is not applicable to environmental problems associated with more reactive pollutants such
*"Guideline on Air Quality Models (Revised)" (1986) and its Supplements; see
reference 3.
-------
as nitrogen oxides (NOJ or ozone. Further, the protocol is oriented towards quantifying
impacts on a source-by-source basis as opposed to assessing the impacts of wide-scale
regional emissions. It has been structured such that both short-term (i.e., averaging times
of 24 hours or less) and long-term (i.e., annual averages) impacts can be determined for
distances in the range of 50 to 300-400km from a source or group of sources. As such,
the procedures outlined here are most likely applicable to regulatory problems dealing
with the prevention of significant deterioration (PSD) or other air quality analyses related
to the development or revision of a State Implementation Plan (SIP) concerning an
individual source or a small group of sources. This document is limited to discussions
of modeling issues, however, and does not address administrative aspects of those
regulatory programs.
The remainder of this document is divided into three chapters. Chapter 2.0 contains
background information on the role of long range transport models in regulatory
applications, and includes a brief discussion of some of the technical aspects of
MESOPUFF II. Those familiar with MESOPUFF II may wish to proceed directly to
Chapter 3.0, which contains the recommended procedures for applying MESOPUFF II
within a regulatory framework. Finally, Chapter 4.0 describes an example problem
illustrating the application of the model to a regulatory situation.
-------
2.0 BACKGROUND
This portion of the protocol contains background information on the role of long
range transport models in regulatory programs and a description of various technical
aspects of MESOPUFF n. As noted in the introductory section, the modeling procedures
described in this document are applicable only to relatively inert pollutants, and are most
applicable to regulatory problems involving PSD and other SIP related analyses. Each
of these are discussed below. That discussion is followed by a brief review of the
theoretical basis of MESOPUFF n, and an overview of the processing steps required to
apply the model.
2.1 ROLE OF LONG RANGE TRANSPORT MODELS
At present, a clear need exists for applying long range transport (LRT) models
to assess the impact of distant SO2 and paniculate matter sources on air quality for
purposes of the PSD program, the preparation of SIPs for nonattainment areas, and the
resolution of interstate transport issues for these pollutants. Typically, assessments of this
type arise through the need to evaluate air quality impacts of existing sources as well as
proposed new sources or modifications to existing sources. With respect to the PSD
program, a number of pristine areas such as National Parks are particularly sensitive to
pollutant levels. Acceptable increases in ambient pollutant concentrations (increments)
have been established for these areas. In some instances, increases in emissions in
upwind areas have already or are soon expected to reduce the available increment for SO2
and/or particulate matter. Because the acceptable increments are relatively small and
because pollutants such as SO2 and particulate matter can be transported over great
distances, the impacts of sources desiring to locate several hundred kilometers upwind
-------
may be significant. In some cases, States are concerned that a major portion of the
available PSD increment may be consumed by sources outside their regulatory
jurisdiction.
Many of the same concerns regarding long range transport of pollutants are
raised by States which must address nonattainment problems for both SO2 and particulate
matter. To the extent that these pollutants are transported long distances, they have in
the past been considered to be part of an "irreducible" background. As more stringent
control programs are implemented, affected States are questioning the effect of
transported pollutants. Section 126 of the Clean Air Act provides States with a
mechanism to require sources affecting nonattainment to be controlled if the impact can
be demonstrated. Thus, modeling techniques are presently needed by States to assess the
contribution of distant sources to SO2 and/or particulate matter nonattainment areas.
In order to address the types of regulatory issues discussed above, LRT
models that have the capability to quantify air quality impacts of sources at distant
locations are required. These types of models must account for plume meander due to
variations in the wind field as well as dispersion of the plume induced by the turbulent
actions of the atmosphere. The modeling requirements for PSD and nonattainment issues
may be further complicated by the need to address transport, dispersion, and coupling of
individual plumes from multiple surface and elevated emission sources. Thus, candidate
LRT models should, at a minimum, have the capability to determine the impacts of more
than one source in a single model evaluation. Nevertheless, this protocol is oriented
towards quantifying impacts on a source-by-source basis as opposed to those more suited
-------
to evaluating wide-scale regional emission impacts (e.g., grid models). Models such as
MESOPUFF n which are point source specific are considered more appropriate for these
types of regulatory applications.
Although a number of mesoscale models have been developed to address the
types of regulatory issues discussed above, they have not been routinely applied to such
problems. Two major impediments to their use are the data requirements and the costs
incurred in applying the models. Another major complicating factor is the need to
consider the impact of source emissions for two distinct averaging periods: short-term
(i.e., 24 hours or less) and long-term (annual). For both averaging periods, NAAQS and
PSD increments are specified such so as to limit the highest ambient concentrations
occurring at a specific location. For example, the SO2 annual average NAAQS
concentration is never to be exceeded at any site, and the short-term 24-hour average
NAAQS concentration may be exceeded only once per year at each location. LRT
models capable of estimating short-term ambient concentrations are relatively expensive
to apply for long time periods, whereas long-term models are more computationally
efficient for long-term simulations but are incapable of estimating short-term
concentrations. Further, techniques have not been developed to select important short-
term periods (i.e., episodes) to model, there-by severely limiting the usefulness of
conducting episodic type analyses to address regulatory problems. The protocol described
in Chapter 3 has been designed to address these problems directly.
-------
2.2 DESCRIPTION OF MESOPUFF H
Presented below is a brief overview of the basic theoretical formulation of
MESOPUFF H. MESOPUFF H is a short-term, plume transport model that
mathematically simulates the transport and dispersion of pollutant emissions from
individual sources. Several preprocessing steps are required to generate the input data
to perform these simulations, and they are described as well. The model also contains
several technical options to account for plume dispersion, dry deposition, chemical
transformation, wet removal, etc. These technical options are discussed in Section 3.4
dealing with the application MESOPUFF H. More detailed discussions of MESOPUFF II
and its preprocessors can be found in reference 1 and 2.
MESOPUFF n is aLagrangian, variable-trajectory, superposition puff model
suitable for simulating the transport and dispersion of air pollutants over distances greater
than about 50km. A continuous plume from a single source is modeled as a series of
discrete puffs (see Figure 2-1). Each puff is transported and dispersed independently, and
ground-level ambient concentrations of a pollutant are calculated at discrete receptors
according to the proximity of the puff to a receptor and the concentration of the pollutant
within the puff. Puffs are emitted at constant, short-term intervals from each source
(e.g., every 15 minutes), and tracked until they leave the user defined modeling domain.
Tracking takes place in two layers, one below the mixing height and the other above.
Transport distance and direction are determined from hourly, gridded wind fields derived
from available meteorological measurements of wind speed and direction. Pollutant
concentrations within a puff vary temporally and spatially as a result of the growth in puff
size due to dispersion. The latter is determined by two sets of time dependent puff
8
-------
Figure 2-1. Schematic Representation of Puff Superposition Model
(adapted from Reference 2)
-------
growth equations relating atmospheric stability to dispersion coefficients, one set for
distances of less than 100km and the other for longer distances.
MESOPUFF II has two associated computer programs for preprocessing key
meteorological measurements in order to generate the gridded data needed for the
transport simulations. The first, READ56, is a preprocessor that edits upper air
rawinsonde measurements for missing information, and produces an output file of the data
in a special format for use by the second preprocessor, MESOPAC II. MESOPAC II
uses these data and hourly surface meteorological data to generate a single output file for
use by MESOPUFF II. Both preprocessors are designed to accept meteorological data
in standard formats as supplied by the National Weather Service (NWS).
Both MESOPAC II and MESOPUFF II employ a Cartesian coordinate
reference frame made up of three nested grid systems: a meteorological grid, a
computational grid, and a sampling grid. The meteorological grid is defined through
inputs to MESOPAC II, and is the basic reference frame for all spatially varying input
data. Thus, coordinates of meteorological stations, sources, and receptors must be
specified relative to this grid. The other two grid systems are subsets of the
meteorological grid. The computational grid defines that portion of the meteorological
grid in which puffs are tracked. The sampling grid can be used to define a group of
gridded receptor points at which ambient concentrations are calculated. Discrete
nongridded receptors can also be used, however, with their coordinates specified relative
to the meteorological grid.
10
-------
MESOPAC n computes meteorological variables such as wind speed, wind
direction, mixing height, stability category, etc., at all nodes of the meteorological grid
for each hour of a simulation to be performed by MESOPUFF n. In addition, the wind
fields are calculated at two levels: a lower level representing boundary layer flow, and
an upper level for flow aloft. The preprocessor uses various spatial and temporal
interpolation schemes that operate on the meteorological measurement inputs. Since LRT
applications may involve transport over relatively long distances, surface and upper air
data from a number of sites near or within the meteorological grid are typically used.
Although MESOPUFF II has the capability to simulate emissions from a
limited number of area sources (up to five), it is primarily point source oriented. Up to
20 individual point sources can be modeled in a single evaluation. As noted above, the
source coordinates are specified relative to the meteorological grid, and the usual other
source data must also be specified (i.e., stack height, stack diameter, effluent exit
velocity, exit temperature, and emission rate).
11
-------
3.0 RECOMMENDED PROCEDURES FOR APPLYING MESOFUFFII
This chapter describes recommended procedures for applying MESOPUFF II and
its preprocessors to regulatory problems associated with the long range transport of
relatively inert pollutants such as SO2 or paniculate matter. As noted earlier, the
procedures recommended in this protocol are most applicable to regulatory problems
involving PSD or other SEP related analyses. These recommendations are general in
nature, and have been developed to foster consistency in applying MESOPUFF II to
problems of these types. They have evolved primarily from results obtained from
conducting model performance evaluations and sensitivity analyses, and have been
developed to be as consistent as possible with general modeling concepts expressed in
"Guideline on Air Quality Models".3 It is recognized, however, that deviations from
these procedures may be warranted in some situations, but in such instances they should
be clearly documented and fully supportable.
The discussion that follows is divided into five broad categories: 1) the spatial and
temporal scales of an analysis, 2) the compilation of a meteorological data base, 3)
application of the MESOPUFF II preprocessors, 4) application of MESOPUFF II, and
5) control strategy evaluation. Each topic is discussed in general terms, with
recommended procedures summarized in a single-spaced format. The discussions that
follow are limited, however, to describing procedures for developing model inputs and
to identifying preferred model options to be used in regulatory applications. Specific
operational aspects of the model and its preprocessors and the formats for coding the
model inputs are described in reference 2.
13
-------
3.1 SPATIAL AND TEMPORAL SCALES OF ANALYSIS
This section deals with defining the modeling region (i.e., the geographical
area of coverage) and selecting the time period for which the model should be applied.
Although the extent of the modeling region must be determined on a case-by-case basis
taking into account the locations of the sources, impact areas, and the meteorological
stations, some general guidelines are presented below. The time period for which the
model is applied determines the length of the model simulations and the period of record
of the meteorological data that is needed for the model application. This aspect of the
model application is dependent on the intended use of the modeling results. Since this
protocol is directed towards PSD and SIP related analyses associated with the NAAQS
and PSD increments, more definitive guidance is provided in this area.
3.1.1 Spatial Scale
MESOPUFF II is applicable to estimating impacts at mesoscale
distances from a source or group of sources. As such, it is most appropriate for dealing
with source impacts in the range of 50 to 400km. Straight-line Gaussian models are
preferable for shorter distances, and regional scale models are more applicable for
transport distances greater than about 400km. In addition, the model is not applicable to
areas with rugged terrain since terrain effects are not directly accounted for in the model.2
The modeling region is defined by the meteorological and
computational grids that were described in Section 2.2. The meteorological grid is the
Cartesian reference frame for all spatially varying input data for both MESOPAC II and
MESOPUFF II. Thus, the locations of all sources, meteorological stations, and receptors
14
-------
must be specified relative to this grid. The SW corner of the grid is represented by the
point (1.0, 1.0). The size of the grid is determined by specifying the number of grid
points in both the west-to-east and south-to-north directions (a maximum of 40 points is
allowed in either direction), and by specifying a common, fixed distance between each
point (i.e., the grid spacing). The computational grid defines the area in which the model
puffs are tracked, and must therefore, encompass both the source locations and the impact
areas. It may be defined to be identical to the meteorological grid or as a subset of that
grid. Figure 3-1 illustrates two hypothetical configurations.
With respect to a particular model application, both the meteorological
and computational grids should be selected to provide broad coverage of the areas of
concern. Since the geographical configuration of the sources and receptors is being
represented by a Cartesian coordinate system (as opposed to one accounting for the
curvature of the earth), the side-widths of the modeling region should probably not
greatly exceed 1000km in the west-to-east direction nor 600km in the south-to-north
direction. The spacing of the grid points may be on the order of 10 to 50km depending
on the overall size of the meteorological grid and the limitation on the number of grid
points that may be used. Since meteorological variables are interpolated at grid points
from measurements made at the meteorological stations, there is little to be gained from
having a much finer resolved spacing of the meteorological grid than the average spacing
between the locations of the meteorological stations. Further, computational savings can
be realized by limiting the resolution of the grid spacing and by restricting the area of
coverage of both the meteorological and computational grids to the immediate area of
concern. For example, if impacts are to be assessed at distances of 50 to 150km
15
-------
Meteorologico! 0"d Computotionoi Grids
Y
1234557
X
EXAMPLE 1
Meteorotogicol
Grid
7
Computationol
Grid
5
Y 4
3
X
234567
X
EXAMPLE 2
Figure 3-1. Example Grid Configurations
16
-------
downwind of the source, no need exists to use a grid system that extends up to 400km
downwind of the source. Nevertheless, sources and receptors should not be located too
near the boundary of the computational grid to avoid possible boundary effects.2 A
cushion of two to three grid points around the edges of the sources and impact areas
should be adequate.
Procedure. MESOPUFF n can be used to estimate source impacts at
distances 50 to 400km downwind. The meteorological and computational grid systems
used to define the modeling region should be formulated so as to encompass all sources
and impact areas, with neither being located too near the edge of the computational grid.
Grid dimensions that greatly exceed 1000km in the west-east direction or 600km in the
south-north direction are not recommended. Finally, spacing of the meteorological grid
on the order of 10 to 50km will probably be adequate for most applications, depending
on the overall grid size and relative spacing between meteorological stations.
3.1.2 Temporal Scale
The second portion of this section deals with the time frames for
conducting the modeling analyses. As described in Section 2.1, PSD and SIP regulatory
analyses typically involve estimating source impacts for both short-term (e.g., 3-hour and
24-hour) and long-term (e.g., annual) averaging periods. Further, MESOPUFF II is
oriented towards evaluating short-term impacts, i.e., it operates on hourly meteorological
data and predicts ambient concentrations at hourly intervals from which concentrations
for longer averaging periods can be computed. As discussed in Section 2.1, procedures
do not currently exist for identifying critical short-term periods a priori (i.e., selecting
those short-term periods with the highest and second-highest concentrations without
running the model for a full year). Thus, it is recommended that the model be applied
for a minimum period of record of one full year. From such an annual simulation, both
the short and long-term critical concentrations can be determined. The minimum 1-year
17
-------
recommendation is made recognizing that computational expense and data availability may
prohibit the routine application of the model to a longer period of record.
In making the recommendation to model a complete year, it is
recognized that computer limitations may preclude completing an annual simulation in a
single model run. Model simulations can be conducted for shorter time periods,
however, and the results concatenated to produce concentration estimates for a complete
year. As described in the next chapter, the annual simulation for the example problem
was developed by performing 12 individual monthly simulations. When this procedure
is used, it will be necessary to set the simulation starting day at least four days prior to
the actual period of interest in order to account for initial transport and dispersion. The
modeling results for the first four days would then be discarded. For example, a
simulation for the month of June would have a starting day of May 28, and the model
predictions for May 28 through May 31 would be ignored. In order to avoid having to
obtain two years of meteorological data to follow this procedure for the beginning of the
year, the starting point of the annual simulation could be January 1 (as opposed to
December 28 of the preceding calendar year).
Recommended Procedure. For regulatory applications of MESOPUFF II, it is
recommended that a minimum one year period of record be simulated in order to identify
the critical short- and long-term impacts. Computational limitations will likely necessitate
that full annual simulations be obtained by performing a series of simulations for shorter
time periods (e.g., monthly periods) and concatenating the model predictions. In such
cases, the starting point for the shorter simulations should be set to four days prior to the
start day for the period being modeled, and the model predictions for these first four days
be discarded. It is not necessary to follow this procedure for the start of the annual
period however.
18
-------
3.2 COMPILATION OF METEOROLOGICAL DATA BASES
The principal meteorological data required for application of MESOPUFF II
include twice-daily upper air soundings and hourly surface meteorological files. The
MESOPUFF II preprocessors are designed to accept data in formats previously used by
the National Climatic Data Center (NCDC) in Asheville, NC. For the upper air data that
format is Standard Tape Deck Format 5600 (TDF5600), and for the surface data it is
Card Deck 144 (CD 144). While the formats currently used by NCDC for archiving these
two data types have changed, NCDC has the capability to convert the new formats to
TDF5600 and CD 144 upon special request. As indicated in the preceding section, at least
one full year of data is required for each meteorological station used in the analysis.
The MESOPAC II preprocessor is designed to handle data from up to 25
surface stations and 10 upper air stations. Since the use of a large number of sites
provides greater confidence in accurately simulating prevailing wind fields and resultant
plume transport, the inclusion of as many stations as possible in the analysis is desirable.
Thus, the use of all NWS sites with available data within or near the edges of the
meteorological grid is recommended. Additional stations surrounding the domain (e.g.,
outside the boundaries) may be included as well if data are available. Should the total
number of stations exceed the maximum allowable for either the upper air or surface
stations, the stations should be selected for the application on the basis of providing the
best spatial coverage throughout the meteorological grid. Since a minimum one year
period of record is needed for each station, it is recommended that the year with the most
available meteorological data be selected for regulatory applications.
19
-------
The MESOPAC n preprocessor requires complete data sets for both TDF5600
upper air data and CD144 surface data (i.e., it is not designed to handle missing data).
In order to provide as much spatial coverage as possible in model applications, it is
recommended that stations not be eliminated on the basis of some missing data. Rather,
the data sets for each station should be screened to identify missing or spurious values,
and then edited to eliminate gaps or clearly erroneous values. The upper air
measurements used by MESOPUFF II include pressure, height, temperature, wind
direction, and wind speed. If a mandatory pressure level (850mb or 700mb) is missing,
it is recommended that the entire sounding from the station be replaced with one from
another station most closely representative of the one with the missing data. Missing
temperatures, wind speeds, and wind directions for any pressure level may be replaced
by an interpolated value using measurements from adjacent levels. The hourly CD 144
data used by the model include cloud cover, ceiling height, precipitation type, wind
speed, wind direction, surface pressure, and temperature. Missing or obviously spurious
values for short time periods may be replaced by interpolating values from adjacent
hours, or assuming an earlier value persists over the period in question. Large data gaps
(e.g., several hours or more) may be replaced by data from another representative station
as a last resort to obtain a complete data base. As described in the next section, READ56
can be used to identify missing upper air data, but it will be necessary for the user to
develop the screening procedure for the surface data.
Finally, the MESOPAC II program is also designed to accept hourly
precipitation data in Tape Deck 9657 format. These data are ultimately used by
MESOPUFF II in estimating the rate of wet removal of pollutants. As will be discussed
20
-------
in Section 3.4, including wet removal in regulatory applications is not currently
recommended. Thus, for the applications discussed here, it is not necessary to obtain or
process these data.
Recommended Procedure. Upper air and hourly surface data from as many stations as
possible that are located within the modeling domain should be included in the modeling
simulations. It is recommended that the year of record be chosen on the basis of
maximum meteorological data availability. Both upper air and surface data sets should
be screened and edited so as to provide complete data sets for the period of record to be
modeled.
3.3 APPLICATION OF MESOPUFF H PREPROCESSORS
The preceding sections described the spatial and temporal scales for applying
MESOPUFF II to regulatory problems, and discussed the meteorological data base that
is needed to perform such applications. This section contains recommendations for using
the meteorological data base with the two MESOPUFF II preprocessors to generate the
information used by the model in a simulation. As described in Section 2.2, the READ56
preprocessor is used to screen upper air data and to produce output files for use by the
second preprocessor, MESOPAC II. MESOPAC II uses the upper air data and the
CD 144 hourly surface data along with other information to construct the temporally and
spatially varying fields of meteorological data used by MESOPUFF II. The use of each
preprocessor is discussed separately below. In these discussions and the ones that follow,
reference is made to some of the variable names used in the MESOPUFF II User's
Manual, and these are shown in capital letters.
21
-------
3.3.1 Application of READ56
As described in Section 3.2, the upper air data need to be screened and
edited prior to using them with the MESOPAC n preprocessor. The READ56 program
can be used to identify missing values and data gaps, but it does not contain built-in
procedures for correcting erroneous or missing data (i.e., the user must make any
necessary corrections to the data files independently of the execution of the program).
Corrections can be made either to the original TDF5600 data, or can be made directly
to the output produced by READ56. Of course, if the first option is chosen it will be
necessary to run the TDF5600 data back through the READ56 program to produce a
corrected output file for use by MESOPAC II. Note that data from only one station can
be processed in a single application of the READ56 program, so it will be necessary to
apply the program separately to each TDF5600 data set used in the analysis.
The input data for READ56 include six variables to control the time
period of the data selected for the run, one variable to define the top pressure level for
each sounding for which data are to be extracted, and four variables to determine how
missing data are handled. Since it is likely that the MESOPUFF II application will be
divided into a number of smaller runs, the first four variables may be used to extract the
TDF5600 data that correspond to the time period of the simulation. The variable for the
top pressure level has to be equal to one of the mandatory pressure levels (i.e., 850, 700,
or 500mb), and it controls the height up to which data are extracted for use in
MESOPAC II. As will be discussed in Section 3.3.2, upper air data are only needed up
to the 700mb pressure level for most applications, so the variable defining the top
pressure level (PSTOP) can be set to 700mb. The remaining four variables control the
22
-------
treatment of missing height, temperature, wind direction, or wind speed found in any one
pressure level. To be consistent with the recommended procedures of Section 3.2, the
control variable for height should be set so that the pressure level is eliminated if the
height is missing (i.e., LHT set to TRUE), and the remaining variables set to flag
missing data (i.e., LTEMP, LWS, and LWD set to FALSE). Any data flagged as
missing must be replaced. Also, mandatory pressure levels (e.g., 850mb and 700mb) that
are missing or eliminated due to missing data will also be flagged and must be corrected.
In both cases the procedures outlined in Section 3.2 may be used.
Recommended Procedure. READ56 may be run for each upper air station to identify
missing data. Data up to 700mb need only be extracted, and the variables controlling
missing data should be set to eliminate a pressure level if the height is missing and to flag
missing temperatures, wind speeds, and wind directions. The flagged data can be
replaced using the procedures outlined in Section 3.2. Finally, corrected READ56 output
files must be generated for use with MESOPAC II.
3.3.2 Application of MESOPAC II
The MESOPAC II preprocessor uses the output files generated by
READ56, the CD 144 surface meteorological data, and other inputs to produce a single
output file for use in a MESOPUFF II simulation. The output file contains time and
space interpolated fields of the following variables: lower and upper level u,v wind
components, mixing height, convective velocity scale, friction velocity, Monin-Obukhov
length, and PGT (Pasquill-Gifford-Turner) stability class. In addition, it contains other
information such as surface roughness length at each grid point, land use category at each
grid point, average surface air density, air temperature at each surface station, etc. To
generate these outputs, MESOPAC II uses several inputs in addition to the basic surface
and upper air data. These inputs have been grouped into 15 different categories in the
23
-------
User's Manual as shown in Table 3-1. Recommended procedures for formulating the
inputs are discussed below by input category. In addition to these recommendations, a
modification to the MESOPAC n program is also suggested, and it is described
immediately following the discussion of the model inputs.
Groups 1-3. Most of the input variables in these groups are simple run
control parameters that will need to be developed on a case-by-case basis (e.g., title,
starting time by number of hours to be processed, number of surface stations, etc.). The
input variables in Group 3 define the meteorological grid, so they should be developed
in accordance with the recommended procedures outlined in Section 3.1.
Group 4. Input variables for this group control the amount and form
of the output produced by MESOPAC n. One variable, LSAVE, controls whether the
output for MESOPUFF II is to be saved to tape or disk. Thus, whenever a
MESOPUFF II simulation is to follow the application of MESOPAC II, this variable must
be set to TRUE, and the output stored on an appropriate storage medium. The remaining
variables control printed output of meteorological data both input to and computed by the
program. Since the volume of printed output from MESOPAC II for a run of any length
can be quite large, these variables will normally be set to suppress printing of the output
results.
Group 5. Land use categories at each grid point of the meteorological
grid must be supplied to MESOPAC II. Table 3-2 shows the 12 different categories that
have been defined for use in MESOPAC II. Appropriate classifications for each grid
24
-------
Table 3-1
Summary of MESOPAC H Run Control Inputs
Group Number Description
1 Run title
2 General information (starting time, number of hours in the run
number of meteorological stations, etc.)
3 Grid data (number of west-east grid points, number of south-north
grid points, grid spacing)
4 Output options (e.g., print control keys)
5 Land use categories for each grid point
6 Default override options
7 Wind speed measurement height (optional)
8 von Karman constant (optional)
9 Friction velocity constants (optional)
10 Mixing height constants (optional)
11 Wind field variables (optional)
12 Surface roughness lengths (optional)
13 Radiation reduction factors (optional)
14 Surface station information
15 Upper air station information
25
-------
Table 3-2
Land Use Categories Used in MESOPAC H*
Category Land Use Type
1 Cropland and Pasture
2 Cropland, woodland and grazing land
3 Irrigated crops
4 Grazed forest and woodland
5 Ungrazed forest and woodland
6 Subhumid grassland and semiarid grazing land
7 Open woodland grazed
8 Desert shrubland
9 Swamp
10 Marshland
11 Metropolitan city
12 Lake or ocean
'adapted from reference 1
26
-------
point must be determined on a case-by-case basis using information obtained from land
use maps or digitized land use inventories.2
Groups 6-13. Group 6 of this set of inputs provides nine options for
overriding default procedures used in calculating various output meteorological variables.
For regulatory applications of MESOPAC II, however, the default procedures are
recommended in all cases. The defaults, which are summarized in Table 3-3, are
consistent with the procedures used in the model performance evaluations conducted with
MESOPUFF n. With the use of all default procedures, no inputs need be supplied for
groups 7 through 13. A tenth option is contained in group 6, however, to indicate
whether the starting point of the MESOPAC II simulation coincides with the starting point
of the meteorological data bases. This option should be set appropriately.
Group 14-15. These two groups of inputs provide information to the
program on the surface and upper air stations such as station identification number, x,y
coordinates, latitude, longitude, and time zone. These data must be determined on a
case-by-case basis. Recall from the discussion in the previous section that hourly
precipitation data are not required for regulatory applications, so corresponding input data
are not needed.
Program Modification. A program modification is recommended for
MESOPAC II to correct a potential problem that could occur infrequently. Under some
meteorological conditions, the mixing height due to mechanical turbulence exceeds the
height corresponding to the 700mb pressure level. When this happens, the program is
27
-------
Table 3-3
Summary of Default Procedures Recommended for
Regulatory Applications of MESOPAC H
1) Wind speed measurement height is equal to 10m.
2) von Karman constant is set to 0.4.
3) Constants used in calculation of friction velocity are set to 4.7 for gamma and
1100 for A.
4) Constants used in calculation of mixing heights are set to 1.41 for B, 0.15 for E,
200m for the layer depth used in estimating the previous hours lapse rate,
0.001° K/meter for the minimum potential lapse rate, and 2400 for the constant N
in the stable (mechanical) mixing height equation.
5) Wind field control variables are set to 99 grid units for RADIUS, 2 for HWF
(i.e., use vertically averaged winds from the earth's surface to the mixing height
for the level wind field), and 4 for IUWF (i.e., use vertically averaged winds
from the mixing height up to the height that corresponds to the 700mb pressure
level for the upper level wind field).
6) Default surface roughness lengths are determined according to specified land use
categories.
7) Default solar radiation reduction factors are based on tenths of cloud cover as
follows:
Cloud Cover (tenths): 0123456789 10
Reduction Factor: 1.0 0.91 0.84 0.79 0.75 0.72 0.68 0.62 0.53 0.41 0.23
8) Heat flux constant at each grid point is set to 0.3.
28
-------
stopped since the average wind flow for the upper layer cannot be computed. The
recommended program modification to correct this problem entails limiting the height of
the mixed layer to 4000m (or to the height of the 700mb level if lower than 4000m) in
the calculations of the average winds aloft. It can be implemented in the program by
modifying subroutine VERTAV as follows:
HTMIX=ZI(IG,JG)
HTMDC=AMIN1(HTMIX,4000.)
ISTAB=IPGT(IG,JG)
where the middle line of code shown above has been added to the subroutine.
Alternatively, another option may be chosen to define the upper level for flow aloft, e.g.,
500mb instead of 700mb (see Section 2.2.1 of the User's Guide2).
Recommended Procedure. All inputs related to run control, meteorological station
identification, meteorological grid definition, and input/output control must be formulated
on a case-by-case basis. Land use categories for each grid point must be defined using
available information according to the classification scheme shown in Table 3.2. Default
values are recommended for all technical options contained in input Group 6 (i.e.,
IOPTS(1), IOPTS(2) ... IOPTS(9) should be set to zero). Finally, it is recommended that
the program be modified as described above in order to limit the height of the mixed
layer to 4000m in the calculation of the vertically averaged winds aloft.
3.4 APPLICATION OF MESOPUFF II
This section contains recommended procedures for developing the model
inputs required to use MESOPUFF II in regulatory applications involving the transport
and dispersion of relatively inert pollutants. As described in Section 2.2, pollutant
emissions from individual sources are modeled as a series of discrete puffs emitted at
regularly spaced time intervals. Ambient concentrations of pollutants are calculated at
29
-------
individual receptor points that can be specified in one of two ways: as part of a gridded
sampling network, or as discrete nongridded points. With the procedure that is
recommended under this protocol, MESOPUFF n is used to compute one-hour average
pollutant concentrations at each receptor for each hour of the simulation. The results are
saved on a computer output file, and later used to calculate concentrations for longer
averaging periods (e.g., 3 hours, 24 hours and annual). This last procedure is discussed
further in Section 3.5.
As noted above, two types of receptor networks can be specified for
MESOPUFF II applications, and both types may be used in the same run. The particular
network(s) that is chosen for any one application will necessarily have to be determined
on a case-by-case basis depending on the purpose of the application. For example, if the
application is intended to assess impacts at a remote Class I PSD area alone, receptor
locations would normally be restricted to the proximity of the area of concern. If the
intent of the application is to identify maximum impacts of a source at distances greater
than 50km downwind regardless of where they occur, then the receptor network would
have to cover a much broader area. Because it is not possible to identify critical short-
term periods without first running the model, it will not be possible to perform screening
tests to identify areas with potentially high concentrations. Although exceptions may
occur, higher concentrations tend to be found nearer the source in long range transport
problems. Thus, a polar coordinate network may be more suited for evaluating source
impacts over large areas since the receptors are more closely spaced along rings nearer
the source. Conversely, a rectangular gridded network may be more appropriate for
assessing impacts in well defined, limited areas. Examples of both are illustrated in the
30
-------
example problem described in the next chapter. Although the use of a large number of
receptors may be desirable to obtain good spatial coverage, it should be emphasized that
computational requirements increase with the number of receptors used.
MESOPUFF n was originally designed to simulate the transport, dispersion,
transformation/formation, and removal of up to five specific individual species: SO2,
SO4=, NOX, HNO3, and NO3~. For the regulatory applications covered by this protocol,
however, it is recommended that the pollutant transformation and removal mechanisms
currently incorporated in MESOPUFF n not be used. Considerable research is underway
to develop techniques for quantifying the effects of these phenomena, but no single set
of approaches has yet gained universal acceptance. As a consequence, it is recommended
that emissions of relatively nonreactive pollutants such as SO2 and particulate matter be
modeled as if they do not react nor are removed from the atmosphere over the transport
distances for which MESOPUFF II is applicable. Sensitivity tests conducted with
MESOPUFF II indicate that including chemical transformation and dry removal in
simulations of SO2 lowers the highest and second highest concentration by about 20 to 30
percent over distances of 50 to 300km downwind from an elevated source. Nevertheless,
these reductions might be offset by considering a longer period of record in an analysis
(i.e., conducting multi-year simulations). Until such time as transformation and removal
processes become better understood and approaches for quantifying their effects are
agreed upon, the most viable approach for dealing with relatively inert pollutants in a
regulatory framework is to assume that plume mass is conserved.
31
-------
Besides the input file that is generated by the MESOPAC n preprocessor,
MESOPUFF n requires a number of other input variables to control the run, set
computational parameters, select various technical options, and define the array of
receptor points used in the simulation. The MESOPUFF n User's Manual categorizes
these inputs according to the 16 groups shown in Table 3-4. Recommended procedures
for developing the inputs are described below.
Groups 1-2. Most of the input variables in these two groups are of the run
control variety, and must be specified on a case-by-case basis (e.g., title, starting day,
starting hour, number of hours in a simulation, etc.). One variable controls the number
of pollutants and their designation. When the transformation and removal mechanisms
are not included in a simulation (as is recommended), however, the pollutant naming
scheme is not relevant since all pollutants are treated identically. Thus, to model only
one pollutant (e.g., SO2 or particulate matter), the number of pollutants (NSPEC) would
be set to 1 and all concentration output produced by MESOPUFF II would be labeled as
if it were SO2. If two pollutants were to be modeled in a single run (e.g., SO2 and
particulate matter), the number of pollutants would be set to 2, and the concentration
outputs for the first and second pollutants would be labeled SO2 and SO4=, respectively.
Group 3. This group consists of seven variables that affect various
computational aspects used in a model simulation. The first controls the averaging time
for the computed concentrations. As discussed above, a one hour period is recommended
(i.e., IAVG=1). Four variables (NPUF, NSAMAD, LVSAMP, and WSAMP) control
the number of puffs released from each source during each hour of a simulation and the
32
-------
Table 3-4
Summary of MESOPUFF H Run Control Inputs
Group Number Description
1 Run title
2 General information (starting time, number of hours in the run,
number of point sources, number of non-gridded receptors etc.)
3 Computational variables (averaging time, puff release rate,
minimum sampling rate, etc.)
4 Grid information (definition of computational and sampling grids)
5 Technical options (vertical concentration distribution, chemical
transformation, dry deposition, etc.)
6 Output options
7 Default override options
8 Dispersion parameters (optional)
9 Vertical diffusivity constants (optional)
10 SO2 canopy resistances (optional)
11 Other dry deposition constants (optional)
12 Wet removal parameters (optional)
13 Chemical parameters (optional)
14 Point source data
15 Area source data
16 Non-gridded receptor coordinates
33
-------
rate at which puffs are sampled in the concentration calculations. In general, accurate
representation of a continuous plume is enhanced by increasing puff release rate and
sampling frequency, albeit at the expense of increasing computational burden. The
recommendations made here were determined from several sensitivity tests designed to
assess the effect of these variables on design concentrations (i.e., high and second high)
at distances varying from 50 to 300km downwind of an elevated source. The results
suggest that a puff release rate of four (NPUF=4) and a minimum sampling rate of two
(NSAMAD=2) represent a reasonable compromise between computational accuracy and
computer resource requirements. In addition, it is recommended that the variable
sampling rate be used (LVSAMP=TRUE), and that the reference wind speed used with
the variable sampling option be 2 m/s (WSAMP=2). The remaining two variables
control the use of a sampling grid and the minimum age for puffs to be sampled. While
the selection of sampling grid must be determined on a case-by-case basis, it is
recommended that the minimum puff sampling age be set to 900s. Since it is
recommended that MESOPUFF II only be used to estimate concentrations at downwind
distances of 50km and beyond, the choice of the latter input is probably not too critical
since its primary purpose is to minimize the possibility of abnormally high concentration
spikes close to the sources (i.e., distances within one hour's travel time).
Group 4. The input variables in this group control the definition of the
computational grid and the sampling grid (if used). As discussed in Section 3.1, the
computational grid can be defined identically to the meteorological grid or as a subset of
that grid. Either approach can be followed, provided that the computational grid
encompasses both the sources and the impact areas, with neither located too near the
34
-------
boundary of the grid. The sampling grid can be used to define an array of rectangular
gridded receptors, and must be determined on a case-by-case basis. It too is determined
relative to the meteorological grid, although a finer spatial resolution for the sampling
grid can be used. The upper limit on the sampling grid is 40x40 points in either
direction.
Group 5. The variables in this group control the use of various technical
options incorporated in the model. As discussed at the beginning of this section, the
recommended approach for applying MESOPUFF n to the types of regulatory situations
considered in this protocol consists of ignoring the potential effects of chemical
transformation, dry deposition, and wet removal. Thus, the three variables controlling
these options (LCHEM, LDRY, and LWET) would be set to omit these processes. One
of the other variables (LGAUSS) in this group controls whether a puff introduced into the
mixed layer is instantaneously dispersed or assumes a Gaussian concentration distribution
in the vertical. Although the significance of this assumption is probably not too great for
travel times greater than a few hours due to plume growth in the vertical direction, the
use of the Gaussian distribution is recommended, primarily to be consistent with the
procedures used in the model performance evaluations. The final variable in this input
group (L3VL) can be used to select three vertical layers when considering dry deposition.
Since the inclusion of dry removal is not recommended, it is recommended that this
option not be used as well.
Group 6. These variables control the output produced by MESOPUFF II.
It will almost always be necessary to save the hourly concentration estimates produced
35
-------
by MESOPUFF H on a tape or disk file by setting the value of LSAVE to TRUE.
Because of the potentially large volume of printed output that can be generated in a fairly
lengthy run of MESOPUFF n, it will usually be desirable to suppress printed output
results using the other print control variables.
Groups 7-13. Group 7 of this set of inputs controls six options for overriding
MESOPUFF n default technical procedures. The first of these can be used to override
the default dispersion parameters. This option is not recommended since the default
technique was used in the model performance evaluation (i.e., IOPTS(1)=0). The
remaining five variables, IOPTS(2) through IOPTS(6), control various technical options
associated with the use of chemical transformation, dry deposition, and wet removal.
Thus, these options would not be invoked under the procedures recommended here (i.e.,
all would be set to zero). The remaining groups (i.e., groups 8 through 13) are used to
input information required to override defaults when so indicated by the override options
selected with the group 7 inputs. Thus, under the procedures recommended here, inputs
would not be required for these six groups.
Groups 14-16. Data inputs for groups 14 and 15 describe the source
characteristics for point and area sources, respectively. Since the protocol is directed
towards point source problems, area source would not typically be included. The point
source data include location, stack height, stack diameter, exit velocity, temperature, and
emission rate. Source data should be developed on a case-by-case basis using annual
average values. If two or more sources are located in close proximity, computational
savings can be realized by combining the sources into one representative source.
36
-------
Reference 5 contains recommended procedures for deriving appropriate stack parameters
for the representative source. Finally, the x and y coordinates (relative to the
meteorological grid) for the nongridded receptors, if any, are input with the last group
of variables. Again, these inputs will typically be determined on a case-by-case basis.
Recommended Procedure. Many of the inputs for MESOPUFF II applications must be
determined on a case-by-case basis (e.g., receptor network specification, number of
species, number of sources, etc.). Specific recommended procedures include: 1)
computing one hour averages; 2) setting the puff release rate to four; 3) setting the
minimum sampling rate to two and using the variable sampling option with a reference
wind speed of 2 m/s; 4) setting the minimum age for puff sampling to 900s; 5) using an
initial Gaussian distribution of puffs in the vertical; 6) not including chemical
transformation, dry deposition, or wet removal in simulations; 7) not using the three
vertical layer option; and 8) using all other default options. Specific inputs corresponding
to these selections are listed below, and key assumptions associated with all
recommendations are summarized in Table 3-5.
IAVG=1
NPUF=4
NSAMAD=2
LVS AMP=TRUE
WSAMP=2.
AGEMIN=900
LGAUSS=TRUE
LCHEM=FALSE
LDRY=FALSE
LWET=FALSE
L3VL=FALSE
IOPTS(1)=0
IOPTS(2)=0
IOPTS(6)=0
37
-------
Table 3-5
Summary of Default Procedures Recommended for
Regulatory Applications of MESOPUFF U
1) One hour average concentrations are computed.
2) Gaussian vertical concentration distribution is assumed for each puff introduced
into the mixed layer.
3) For distances up to 100km, the dispersion parameters are from functions fitted to
the curves of Turner.6 For longer travel distances, time dependent growth
functions from Heffter are used.7
4) Growth rates for puffs above the mixed layer are those corresponding to
E stability.
5) Chemical transformation, dry deposition, and wet removal processes are not
included in the simulations.
6) The three vertical layer option is not used.
7) Four puffs are released from a source each hour, variable sampling rates are used
depending on wind speed, and no puff is sampled within the first 900 seconds of
its release.
38
-------
3.5 CONTROL STRATEGY EVALUATION
This section contains a brief generalized discussion of how the MESOPUFF II
modeling results may be processed to assess air quality impacts, and if necessary,
determine appropriate emission limitations for a source or group of sources. This
protocol is directed towards SIP and PSD analyses associated with SO2 and paniculate
matter, so the air quality measures of most concern are the NAAQS and the PSD
increments established for these pollutants. Since these measures all involve averaging
times longer than one-hour, the discussion below begins with a description of how the one
hour average concentrations generated by MESOPUFF II can be used to compute
concentration estimates for longer averaging periods.
When the recommended procedure for applying MESOPUFF II to regulatory
problems is followed, MESOPUFF II produces an output file which contains estimated
one-hour average concentrations at each receptor. As described earlier, it will most likely
be necessary to divide a full annual simulation into a series of shorter simulations due to
computer limitations. To facilitate further processing of that data, it will probably be
advantageous to combine the results obtained from all simulations into a single file
containing the concentration estimates for the full year. From this file it is possible to
compute concentrations for longer averaging periods and determine the highest and
second-highest values predicted to occur during a year that are needed for comparison
with the relevant NAAQS or PSD increment. When calculating 3- and 24-hour averages,
it is recommended that nonoverlapping periods be used in the calculations (i.e., block
averages as opposed to running averages). Annual averages should be computed by
summing the hourly concentrations and dividing by the number of hours for which
39
-------
concentrations are available. These calculations can be performed by using the
MESOPUFF II postprocessor (MESOFILE) or software developed by the user.
The concentration estimates output by MESOPUFF n represent the total
impacts of all sources in a simulation, and the program is not designed to produce a
source contribution file. Thus, some special considerations may be required for
evaluating the effects of lowering (or raising) emissions from an established base case
(e.g., evaluating a control strategy). When only one source has been included in a
simulation, the estimated air quality concentrations are directly proportional to the source
emission rate. Thus, the effects of changes in emissions can be evaluated directly without
rerunning the model. Further, if two pollutants are being modeled in this situation, some
computer time could be saved by modeling only one pollutant and deriving the results for
the other by scaling according to the ratio of the emission rates.
Control strategy evaluation is more complicated when multiple sources are
included in a simulation. If an estimated concentration exceeds some acceptable limit,
a control strategy could always be evaluated by changing source emission rates and
rerunning the model for the entire year. This can be relatively expensive, however,
especially if a large number of strategies are to be evaluated. Some savings may be
realized if only a few, short-term episodes are identified as needing additional evaluation
(assuming the emission rate at any source is not increased). In this case, only the critical
short-term periods would need to be remodeled, but it would be necessary to begin the
simulation at least four days prior to the episode of interest. If a large number of
episodes are found in which a concentration estimate exceeds an acceptable value,
40
-------
however, it may be easier to rerun the full year with new emission rates. Of course, the
full year would need to be rerun if a predicted annual average concentration exceeded an
acceptable limit.
Recommended Procedure. Concentration estimates for averaging periods longer than
one hour (e.g., 3 hours and 24 hours) should be computed as nonoverlapping (i.e., block)
averages. Annual averages concentrations should be computed using all hourly estimates.
If only one source is included in a simulation, predicted concentrations are directly
proportional to emission rates, and control strategies can be evaluated directly without
rerunning a full annual simulation. If multiple sources are included in a simulation,
control strategy evaluations must be carried out using MESOPUFF n to evaluate all
critical short-term periods, either by simulating a full year or by modeling episodes only
if unacceptably high concentrations are found for short-term periods alone.
41
-------
4.0 EXAMPLE MESOPUFF n APPLICATION
This chapter illustrates the application of MESOPUFF n and its preprocessors to
an example regulatory problem using the procedures recommended in the preceding
chapter. The example application consists of quantifying air quality impacts of a
hypothetical proposed new source at distances of 50 to 300km from the source. This
example has been developed for illustrative purposes only, and is not intended to
represent an existing or pending regulatory issue involving any specific source. Although
the example is presented as one involving emissions of SO2, the procedures described
below could just as easily have been applied to a similar problem dealing with particulate
matter.
The discussion below is divided into four sections. The first contains a general
description of the example problem. The next section describes the meteorological data
base used in the analysis, and the application of the MESOPUFF II preprocessors. The
final two sections discuss the MESOPUFF II application and summarize the results of the
impact assessment.
4.1 DESCRIPTION OF EXAMPLE PROBLEM
The example problem involves modeling SO2 emissions from a single power
plant stack located near the Ohio River in southeastern Ohio. The modeling region, as
defined by the meteorological grid, was chosen to encompass the source and the impact
areas, and to make use of meteorological data that are available for this region of the
country. Figure 4-1 shows the area encompassed by the boundaries of the meteorological
43
-------
MODELING REGION
Figure 4-1. Modeling Region and Source Location
44
-------
grid and the location of the source relative to the modeling region, and Table 4-1 lists the
relevant source data used for the example problem.
As noted above, the modeling problem entails assessing the air quality impacts
of SO2 emissions at distances of 50 to 300km downwind from the source. Two distinct
types of assessments are considered. In the first, regionwide air quality impacts in the
50 to 300km range are included in the analysis, regardless of location. Of primary
interest are the model predictions of the highest and second-highest ambient
concentrations for several different averaging times. This information might be used, for
example, to assess the impact of a proposed new source on attainment of the NAAQS.
The second type of assessment involves determining impacts in a limited, predefined area
located approximately 200km from the source. This type of application illustrates how
MESOPUFF II might be used in a PSD problem involving the consumption of available
increments in a Class I or II area. As with the first type of assessment, impacts for
multiple averaging times are considered. Both assessments were performed in the same
MESOPUFF II application in which one full year was simulated, the minimum period of
record that is recommended for regulatory applications.
4.2 PREPROCESSOR APPLICATIONS
As described in the preceding portions of this document, the modeling region
is defined by means of the meteorological grid shown in Figure 4-2. This grid system
consists of 24 points in the west-east direction and 21 points in the south-north direction,
with a grid spacing of 40km between each point. For most applications of this type, the
meteorological grid would normally be defined such that the source is located near the
45
-------
Table 4-1
Source Data For Example Problem
Stack Height 250m
Stack Diameter 8m
Exit Velocity 26 m/s
Stack Gas Temperature 430° K
Emission Rate 6000 g/s
46
-------
21
20
1S
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
S
2
1 1 1
I 1 1
~r ~f~i
1 S ' -
r r r
i i i
i j i
L L L J
1 1 1
1 1 1
-r-r-r-
i i t
TTT*"
i i i
i i i
i i i
i i i
-r-r-r-
i i i
r rji "
i i i
i i i
._L_L_L-
! 1 1
f -J--I--
1 1 1
-r -r-r-
i i i
~r~r~r~
i i i
i i f
l ! 1
c U
I l I i wi i i
111111!$
~~1~*1~T~TT~r"~J
IIIIIII
1 1 1 1 1 t 1
J_ J A Ci-L L L
IIIIIII
_ J _ -J .1 .1 i-ĞU I .-
IIIIIII
1 1 1 1 1 ^1 1
|S i i i i i i
IIIIIII
- J_J_J_1_L_L_L_
1 1 1 1 S 1 1
- -i- -1-4-4- -4--1--1--
IIIIIII
- -I-I--T- t-r-r-r-
1 1 c 1 1 1 1 1
1111111
1111111
_ J-J. J.l.L.L-L-
i I l i i i r
iiiiiii
- i--i --r - T-T-I- -r-
iiiiiii
U i i i i i i i
1111111
_L ! i J 1 1 !
IIIIIII
IIIIIII
IIIIIII
~TTI T~r~r r~
1 Ic 1 1 1 I i
i i i 7 J i i
J J J _1 -L _ L _L _
IIIIIII
i i i i i U i
- 1 - 1 - -i ~ r - r - r - r -
t i i i i i i i
r"]~TT~r~r~r~
1111111
i i ic i i i i
_ J_ J_ J^l_l_L_L-
IIIIIII
--I--I-4-4--4--I--1--
IIIIIII
-fl--i- -t -T -r -r-r-
MIIIIII
1111111
i i i t i i i
.. J_J_ J_1_L_L_L_
I 1 1 1 1 1 1
--1--I-4-4--J--I--J--
IIIIIII
- -i - -i - -r - 1 - r - r - r -
iiiiiii
1 I ! t t 1 Ul
iiiiiii
U
i i
i i
! 1
1 1
1 1
l_ J_
1 1
1 1
~ *1~ "1 "~ "
I I
1 1
1 1
1 I
--I --I-.
1 1
--r 1--
l l
1 I
i l
_ J_ J_.
l l
i i
1 i
l i
i i i
T i
1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24
S = SURFACE U = UPPER AIR AND SURFACE STATION
Figure 4-2. Locations of Surface and Upper Air Stations
47
-------
center of that grid. The particular meteorological grid used in this application is extended
somewhat in the westerly direction, however, in order to include more meteorological
stations in the preprocessor applications. Meteorological data used in the study are from
measurements taken during 1975 at 23 NWS surface stations and 7 NWS upper air
stations. The location and identification numbers of each station are listed in Table 4-2,
and the location of each station relative to the meteorological grid network is shown
schematically in Figure 4-2.
As recommended in section 3.2, all meteorological data were screened and
edited to eliminate missing or spurious values. The upper air data consisted of twice-
daily upper air soundings taken at 00 GMT (7 PM EST) and 12 GMT (7 AM EST) at
each station for the entire year. Missing temperatures, wind speeds, or wind directions
within any one pressure level were replaced with values interpolated from adjacent
pressure levels. A pressure level was eliminated if the height was missing, and an entire
sounding was replaced with one from another station if a mandatory pressure level was
missing. Both procedures were carried out using an automated routine that operated on
the basic upper air data files (i.e., changes were made to base upper air data files as
opposed to the READ56 output files). The replacement of an individual sounding with
one from another station was performed using the replacement order shown in Table 4-3.
Data for each station were processed for the complete year, and then supplied to the
READ56 program for final checking and subsequent creation of the input files used by
MESOPAC II.
48
-------
Table 4-2
List of Surface and Upper Air Stations
Location
Pittsburgh
Erie
Buffalo
Huntington
Parkersburg
Greensboro
Louisville
Nashville
Cleveland
Dayton
Columbus
Cincinnati
Toledo
Detroit
Flint
Grand Rapids
Evansville
Fort Wayne
Indianapolis
South Bend
Youngstown
Chicago
Lansing
Surface Station Number
94823
14860
14733
03860
03804
13723
93821
13897
14820
93815
14821
93814
94830
94847
14826
94860
93817
14827
93819
14848
14852
14819
14836
Upper Air Station Number
72520
72528
72425
72317
72327
72429
72637
49
-------
Table 4-3
Primary and Alternate Upper Air Stations*
Greensboro. NC (12311}
Huntington, WV (72425)
Cape Hatteras, NC (72304)
Athens, GA (72311)
Nashville, TN (72327)
Nashville. TN (72327)
Salem, IL (72433)
Athens, GA (72311)
Huntington, WV (72425)
Greensboro, NC (72317)
Huntington. WV (72425)
Dayton, OH (72425)
Pittsburgh, PA (72520)
Washington, D.C.-Dulles
(72403)
Greensboro, NC (72317)
Dayton. OH (72429)
Huntington, WV (72425)
Pittsburgh, PA (72520)
Flint, MI (72637)
Nashville, TN (72327
Pittsburgh. PA (72520)
Buffalo, NY (72528)
Washington, D.C.-Dulles
(72403)
Huntington, WV (72425)
Dayton, OH (72429)
Buffalo. NY (72528)
Pittsburgh, PA (72520)
Flint, MI (72637)
Albany, NY (72518)
Washington, D.C.-Dulles
(72403)
Flint. MI (72637)
Dayton, OH (72429)
Green Bay, WI (72645)
Buffalo, NY (72528)
Pittsburgh, PA (72520)
'Alternate stations are listed below the primary station in descending order of
preference. NWS station numbers are shown in parentheses.
50
-------
The hourly surface data used for the MESOPAC n applications had been
previously screened and edited for use with straight-line Gaussian models applicable to
distances up to 50km. As part of that procedure, the wind data for hours with calms had
been modified such that the wind speed was set to 1.0 m/s and the wind direction to the
value for the preceding hour. Although this procedure would not necessarily have to be
performed for a MESOPUFF II application, no attempt was made to convert the data
back to its original form. It is unlikely that changing the wind data for calms in this
manner would lead to significant differences in model predictions since the wind fields
are developed by interpolating the wind direction and speed at each grid point using
measurements from several stations.
Besides the meteorological data, the only other set of regional specific inputs
required by MESOPAC II is the land use categories that must be assigned to each grid
point, these values were obtained from available land use maps using the classification
scheme listed in Table 3-2. All other MESOPAC II inputs were chosen in accordance
with the recommended procedures described in Section 3-3.
As indicated above, both the surface and upper air data were screened and
edited to produce computerized data sets for the entire year. Because of computer
limitations, however, it was not practical to generate a single MESOPAC II output file
for the full year. Instead, the model simulations for a complete year were divided into
12 monthly simulations, with each month processed by first running READ56, followed
by MESOPAC H, and finally MESOPUFF II. Only the MESOPUFF II output data were
saved for further processing. For most runs of READ56 and MESOPAC II, the time
51
-------
period for the application was selected such that it overlapped the time period for the
subsequent preprocessor or model run. For example, if the MESOPUFF n simulation
period was to include May 28 through June 30, the MESOPAC n application would be
set to cover May 27 through July 1, and the READ56 period to cover May 26 through
July 2. This approach ensured that the output files generated by each preprocessor
contained a sufficiently lengthy period of record for the application of the next program
in the series. For the beginning and ending periods of the annual simulation, however,
the same periods were used for all program applications. An example input data set for
both READ56 and MESOPAC n are contained in Appendix A. A description of the
computer time necessary to apply these preprocessors is included in the next section.
4.3 APPLICATION OF MESOPUFF H
The MESOPUFF II simulations were conducted using inputs that were
developed according to the procedures outlined in Section 3.4. As described earlier in
this section, two objectives of the example application were to estimate regionwide source
impacts at distances of 50 to 300km, and to estimate the impacts at a predefined area that
might represent a remote Class I or II PSD area. The application of MESOPUFF II to
both types of assessments is described below.
In order to determine regionwide impacts of the source between 50 and 300km
downwind, two receptor networks were used. The first network consists of three
concentric rings located at downwind distances of 50, 150, and 300km from the source,
with receptors spaced at 20° intervals on each ring. This network is designed to detect
air quality impacts in all directions from the source. The second network is a group of
52
-------
more densely packed receptors located along arcs extending from 20° to 110° east of the
source. Downwind arc distances extend from 50 to 300km at incremental distances of
50km, and receptors are positioned at 10° intervals along each arc. With the second
network, more receptors are located in the predominant downwind direction since it is
anticipated that higher concentrations would be found in this area. A schematic
representation of both receptor networks is shown in Figure 4-3A.
The second example assessment consists of determining the source impacts at
a remote Class I or Class n PSD area. For this assessment, it was assumed that such an
area is located approximately 200km to the east-northeast of the source. This area was
covered by a group of uniformly gridded receptors spaced at 20km intervals. Figure
4-3B shows the location of the PSD area relative to the source, and the receptor grid
network used to assess impacts in that area.
All of the MESOPAC II and MESOPUFF II simulations were conducted using
the Sperry 1100 computing system at EPA's National Computing Center (NCC). As
indicated earlier, the year-long simulation was divided into 12 shorter monthly
simulations. Except for the first month, all simulations were started four days prior to
the beginning of the month and the results for those days discarded. The model predicted
concentrations for each month were saved in individual files, and later concatenated into
a single output file containing data for the entire year. Table 4-4 summarizes the
computer resource usage required to apply MESOPUFF II and its preprocessors. The
variations in computer time required to perform the MESOPUFF II simulations reflect
53
-------
RING
ARC
SOURCE
Ğ *
A) Regionwide Assessment
GRID
SOURCE
B) PSD Assessment
Figure 4-3. Schematic Representation of Receptor Networks
54
-------
Table 4-4
Summary of Computer Resources Used for Monthly Simulations
Program Storage* CPU Time*
READ56 20K 2
MESOPAC H 100K 45
MESOPUFF H 100K 60-120
'Words of core storage required
"Minutes of CPU time on the EPA NCC Sperry 1110 computer
""Total CPU time for complete annual simulation (all three programs) is
approximately 25 hours
55
-------
differences in puff residence times and sampling frequency occurring between months.
Example inputs for MESOPUFF n are contained in Appendix A.
4.4 SUMMARY OF RESULTS
Although the results obtained from the example application are specific to the
situation modeled, they can provide some insight into prediction patterns that might be
obtained through a real application. Table 4-5 lists the highest and second-highest
concentrations predicted at the receptors used in the regionwide assessment for four
different averaging times. For reference, the levels of the NAAQS are also shown where
applicable. To provide some indication of where the greatest impacts occur, Table 4-6
was prepared. This table lists the top ten second-high concentrations for the short-term
averages and the ten highest annual average concentrations predicted at the receptors used
in the regionwide assessment. The location of the receptor relative to the source and the
network of which it is a part is also shown for each entry. Recall from previous
discussions that receptors located in the arc network are in the expected predominate
downwind direction, whereas those in the ring network are outside this region. As is
evident from Table 4-6, most of the second-high concentrations for the short-term
averaging periods tend to occur nearest the source (i.e., 50km) and in the predominate
downwind direction. For the longer averaging time period, however, this tendency is not
as great. These results suggest that significant source impacts as predicted by
MESOPUFF II could occur at variable distances downwind from the source, and in
virtually any direction. For this particular application, the predicted concentrations are
well below the level of the SO2 NAAQS. Thus, unless other sources were to contribute
56
-------
Table 4-5
Greatest Regionwide Impacts*
Averaging
Time
1-hr
3-hr
24-hr
annual
High
2nd High
683
350
338
194
79
55
3
3
Downwind
Distance
50km
100km
100km
50km
50km
50km
150km
50km
Level of
NAAOS"
NA
NA
NA
NA
NA
365
80
NA
*A11 concentrations are in units of /zg/m3.
**NA = not applicable.
57
-------
Table 4-6
Top Ten High/Second-high Predicted Concentrations*
Regionwide Assessment
Receptor
Network
Arc
Arc
Arc
Arc
Arc
Ring
Arc
Arc
Ring
Arc
Arc
Arc
Ring
Ring
Arc
Ring
Ring
Ring
Arc
Arc
Arc
Arc
Arc
Arc
Ring
Ring
Arc
Arc
Ring
Ring
Direction from
Source"
1-hour averaging
110
40
110
100
30
160
70
80
280
20
3-hour averaging
110
100
140
240
30
300
260
280
70
80
24-hour averaging
110
40
100
30
240
120
90
70
140
320
Downwind
Distance'"
time
100
50
50
50
50
50
50
50
50
50
time
50
50
50
50
50
50
50
50
50
50
time
50
50
50
50
50
150
50
50
50
50
Second-high
Concentration
350
327
324
308
307
268
262
258
258
257
194
181
181
157
156
154
152
148
142
142
55
50
41
40
37
37
35
31
30
30
Annual averages
Arc
Arc
Arc
Arc
Arc
Arc
Arc
Ring
Arc
Ring
60
60
80
100
80
100
120
120
40
140
150
50
50
50
150
150
50
150
50
50
Highest
Concentration
3
3
3
2
2
2
2
2
2
2
"All concentrations in units of ptg/m3
""compass degrees, with due North equal to 0°
"""kilometers
58
-------
substantially to ambient concentrations in this area, no further control strategy evaluations
would appear warranted for this assessment.
The second portion of the regulatory application involved estimating source
impacts at the hypothetical PSD area that is covered by the gridded receptor network
located approximately 200km downwind. Table 4-7 shows the extremes of the second-
high short-term concentrations and highest long-term concentrations predicted at the
receptors located in this grid. The allowable PSD increments for both a Class I area
(i.e., pristine area) and a Class n area (i.e., an area currently attaining the NAAQS) are
also shown where applicable. The highest and lowest impacts of concern differ by about
a factor or two. Less variation in model predictions is found for this limited area than
for the more extensive ring and arc networks. The highest second-high concentrations
are all significantly lower than the Class II PSD increments, but exceed the allowable
increments for a Class I area. Thus, the modeling results suggest that SO2 emissions
would need to be reduced by about 67 percent to meet the allowable 24-hour average
increment for a Class I area assuming no other source contribution.
The model application described above illustrates how a model such as
MESOPUFF II can be used within a regulatory framework to address two different types
of problems. Although the processing time and the computer costs are not insignificant,
the model can be applied to simulate a yearly record of short-term concentrations for
regulatory assessments (e.g., for comparison with NAAQS or allowable PSD increments).
It should be noted that the example problem included only one source, and greater
computer costs would be incurred if multiple sources were modeled. The costs for
59
-------
Table 4-7
Summary of Model Predictions for PSD Assessment*
Class I Class II
Highest Lowest PSD PSD
Averaging Time 2nd High 2nd High Increment**
Increment**
1 hour 68 30 NA NA
3 hours 40 20 25 512
24 hours 15 7 5 91
annual 2 1 2 20
"All concentrations are in. units of /ig/m3, rounded to the nearest ftg/m3; lowest and highest maxima are shown for
the annual average.
"NA = not applicable.
60
-------
MESOPAC n and READ56 would remain the same, however, so the total costs of
modeling more than one source would not increase proportionally.
61
-------
62
-------
5.0 REFERENCES
1. Environmental Protection Agency, 1984. Development of the MESOPUFF II
Dispersion Model. EPA Contract No. 68-023733, U.S. Environmental Protection
Agency, Research Triangle Park, NC.
2. Environmental Protection Agency, 1984. User's Guide to the MESOPUFF II
Model and Related Preprocessor Programs. EPA Publication No. EPA-600/8-84-
013. U.S. Environmental Protection Agency, Research Triangle Park, NC. (NTIS
No. PB 84-181775)
3. Environmental Protection Agency, 1986. Guideline on Air Quality Models
(Revised) and its Supplements. EPA Publication No. EPA-450/2-78-027R. U.S.
Environmental Protection Agency, Research Triangle Park, NC. (NTIS No.
PB 86-245248)
4. Environmental Protection Agency, 1986. Evaluation of Short-term Long-Range
Transport Models, Volumes I and H. EPA Publication Nos. EPA-450/4-86-016a
and b. U.S. Environmental Protection Agency, Research Triangle Park, NC.
(NTIS Nos. PB 87-142337 and PB 87-142345, respectively)
5. Environmental Protection Agency, 1992. Screening Procedures for Estimating the
Air Quality Impact of Stationary Sources, Revised. EPA Publication No.
EPA-454/R-92-019. U.S. Environmental Protection Agency, Research
Triangle Park, NC.
6. Turner, D. B., 1970. Workbook of Atmospheric Dispersion Estimates, AP-26,
U.S. Environmental Protection Agency, Research Triangle Park, NC.
7. Heffter, J. L., 1965. The Variations of Horizontal Diffusion Parameters with Time
for Travel Periods of One Hour or Longer. Journal of Applied Meteorology,
4: 153-156.
63
-------
APPENDIX A
EXAMPLE INPUT DATA SETS FOR READ56, MESOPAC II
AND MESOPUFFII
A-l
-------
Example Inputs for READ56
75 146 00 75 183 23 500.
F F F
A-2
-------
Example Inputs for MESOPAC II
MESOPAC H SIMULATION FOR JUNE
75 147 864 23
24 21 40000.
T F
2 21212 2 2
1 11212 2 2
1 11212 2 2
1111 2
111122
111122
112222
112222
112222
112222
112222
112222
222222
222222
222222
222222
222222
222244
222244
222222
222222
1
94823 18.91
14860 18.94
14733 21.88
03860 13.86
03804 16.32
13723 19.61
93821 7.52
13897 5.46
14820 15.47
93815 10.70
14821 13.23
93814 9.80
94830 11.41
94847 12.44
14826 11.68
94860 7.78
93817 3.58
14827 8.53
93819 6.29
14848 6.17
14852 17.95
14819 3.19
14836 9.66
72520 18.91
72528 21.88
72425 13.86
72317 19.61
72327 5.46
72429 10.70
72637 11.68
7 5
12 F
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
4
5
5
2
2
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
422
422
555
555
255
255
14.58
18.98
21.39
8.69
11.39
2.34
8.17
2.42
17.08
13.01
13.23
10.66
17.72
19.45
21.51
21.22
7.78
16.00
12.46
17.91
16.71
18.14
20.98
14.58
21.39
8.69
2.34
2.42
13 . 0.1
21.51
2
2
2
2
1
1
2
2
2
2
2
2
2
2
5
5
5
5
5
5
5
2
2
2
1
222
2 2
222
2 2121212121212
2 2121212121212
11212
11212 2
11212 2
2
2
2
2
2
2
2
2
5
5
5
5
5
5
5
222
222
222
222
222
222
222
222
555
222
222
222
222
222
222
40.50
42.08
42.93
38.37
39.35
36.08
38.18
36.12
41.40
39.90
40.00
39.07
41.60
42.23
42.97
42.88
38.05
41.00
39.73
41.70
41.27
41.78
42.78
40.50
42.93
38.37
36.08
36.12
39.90
42.97
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
5 5
2 2
2 2
2 2
2 2
2 2
2 2
222
222
222
222
222
255
255
255
255
255
255
522
222
222
222
222
222
222
80.22
80.18
78.77
82.55
81.43
79.95
85.73
86.68
81.85
84.22
82.88
84.67
83.80
83.33
83.73
82.52
87.53
85.20
86.28
86.32
80.67
87.75
84.60
80.22
78.77
82.55
79.95
86.68
84.22
83.73
2 2
2 2
2 2
5 5
5 5
5 5
2 2
2 2
5 5
5 5
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
9 9
9 9
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.0
5.0
5.0
5.0
5.0
6.0
5.0
5.0
5.0
5.0
5.0
6.0
5.0
5.0
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
61
62
63
64
02
03
04
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
999
A-3
-------
Example Inputs for MESOPUFF U
MESOPUFF H SIMULATION FOR JUNE
75
1
1
T
T
000000
15.6 11
15
16
16
16
16
16
16
16
16
15
15
14
14
14
14
14
14
15
15
16
18
18
19
19
18
18
16
15
14
13
12
11
11
12
13
14
15
18
20
22
22
22
22
20
18
15
13
10
148
4
24
F
F
01 816 1
2 T 2.
1 21 7
F F F
01 F 1
0 139 1
T 900.
20 6 15
24
2
.4 250. 8.00 26.0 430. 6000.
.60
.03
.40
.68
.83
.83
.68
.40
.03
.60
.17
.80
.52
.37
.37
.52
.80
.17
.60
.88
.01
.85
.29
.29
.85
.01
.88
.60
.32
.19
.35
.91
.91
.35
.19
.32
.60
.17
.42
.10
.99
.99
.10
.42
.17
.60
.03
.78
12.
12.
12.
12.
11.
11.
10.
10.
10.
10.
10.
10.
10.
11.
11.
12.
12.
12.
15.
14.
14.
13.
12.
10.
9.
8.
7.
7.
7.
8.
9.
10.
12.
13.
14.
14.
18.
18.
17.
15.
12.
10.
7.
5.
4.
3.
4.
5.
65
57
36
02
62
18
77
44
23
15
23
44
77
18
62
02
36
57
15
92
27
28
05
75
52
53
88
65
88
53
52
75
05
28
27
92
90
45
15
15
70
10
65
65
35
90
35
65
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
150
150
150
150
150
150
150
150
150
150
150
150
150
150
150
150
150
150
300
300
300
300
300
300
300
300
300
300
300
300
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
RING
20
40
60
80
100
120
140
160
180
200
220
240
260
280
300
320
340
0
40
60
80
100
120
140
160
180
200
220
240
260
280
300
320
340
20
40
60
80
100
120
140
160
180
200
220
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
DEG
A-4
-------
Example Inputs for MESOPUFF
(continued)
9.10 7.65 49 300.0 KM RING 240.0 DEG
8.21 10.10 50 300.0 KM RING 260.0 DEG
8.21 12.70 51 300.0 KM RING 280.0 DEG
9.10 15.15 52 300.0 KM RING 300.0 DEG
10.78 17.15 53 300.0 KM RING 320.0 DEG
13.03 18.45 54 300.0 KM RING 340.0 DEG
16.03 12.57 55 50.0 KM ARC 20.0 DEG
16.22 12.48 56 50.0 KM ARC 30.0 DEG
16.40 12.36 57 50.0 KM ARC 40.0 DEG
16.56 12.20 58 50.0 KM ARC 50.0 DEG
16.68 12.02 59 50.0 KM ARC 60.0 DEG
16.77 11.83 60 50.0 KM ARC 70.0 DEG
16.83 11.62 61 50.0 KM ARC 80.0 DEG
16.85 11.40 62 50.0 KM ARC 90.0 DEG
16.83 11.18 63 50.0 KM ARC 100.0 DEG
16.77 10.97 64 50.0 KM ARC 110.0 DEG
16.46 13.75 65 100.0 KM ARC 20.0 DEG
16.85 13.57 66 100.0 KM ARC 30.0 DEG
17,21 13.32 67 100.0 KM ARC 40.0 DEG
17.52 13.01 68 100.0 KM ARC 50.0 DEG
17.77 12.65 69 100.0 KM ARC 60.0 DEG
17.95 12.26 70 100.0 KM ARC 70.0 DEG
18.06 11.83 71 100.0 KM ARC 80.0 DEG
18.10 11.40 72 100.0 KM ARC 90.0 DEG
18.06 10.97 73 100.0 KM ARC 100.0 DEG
17.95 10.54 74 100.0 KM ARC 110.0 DEG
16.88 14.92 75 150.0 KM ARC 20.0 DEG
17.47 14.65 76 150.0 KM ARC 30.0 DEG
18.01 14.27 77 150.0 KM ARC 40.0 DEG
18.47 13.81 78 150.0 KM ARC 50.0 DEG
18.85 13.28 79 150.0 KM ARC 60.0 DEG
19.12 12.68 80 150.0 KM ARC 70.0 DEG
19.29 12.05 81 150.0 KM ARC 80.0 DEG
19.35 11.40 82 150.0 KM ARC 90.0 DEG
19.29 10.75 83 150.0 KM ARC 100.0 DEG
19.12 10.12 84 150.0 KM ARC 110.0 DEG
17.31 16.10 85 200.0 KM ARC 20.0 DEG
18.10 15.73 86 200.0 KM ARC 30.0 DEG
18.81 15.23 87 200.0 KM ARC 40.0 DEG
19.43 14.61 88 200.0 KM ARC 50.0 DEG
19.93 13.90 89 200.0 KM ARC 60.0 DEG
20,30 13.11 90 200.0 KM ARC 70.0 DEG
20.52 12.27 91 200.0 KM ARC 80.0 DEG
20.60 11.40 92 200.0 KM ARC 90.0 DEG
20.52 10.53 93 200.0 KM ARC 100.0 DEG
20.30 9.69 94 200.0 KM ARC 110.0 DEG
17.74 17.27 95 250.0 KM ARC 20.0 DEG
18.72 16.81 96 250.0 KM ARC 30.0 DEG
19.62 16.19 97 250.0 KM ARC 40.0 DEG
20.39 15.42 98 250.0 KM ARC 50.0 DEG
21.01 14.53 99 250.0 KM ARC 60.0 DEG
21.47 13.54 100 250.0 KM ARC 70.0 DEG
21.76 12.49 101 250.0 KM ARC 80.0 DEG
21.85 11.40 102 250.0 KM ARC 90.0 DEG
21.76 10.31 103 250.0 KM ARC 100.0 DEG
21.47 9.26 104 250.0 KM ARC 110.0 DEG
A-5
-------
Example Inputs for MES0POTFII
(continued)
18.
19.
20.
21.
22.
22.
22.
23.
22.
22.
20,
20.
20.
20.
20.
21.
21.
21.
21,
21.
21.
21.
21.
21.
21.
22.
22.
22.
22.
22.
22.
22.
22.
22.
22.
17
35
42
35
10
65
99
10
99
65
60
60
60
60
60
10
10
10
10
10
60
60
60
60
60
10
10
10
10
10
60
60
60
60
60
18
17
17
16
15
13
12
11
10
8
12
13
13
14
14
12
13
13
14
14
12
13
13
14
14
12
13
13
14
14
12
13
13
14
14
.45
.90
.15
.22
.15
.97
.70
,40
.10
.83
.65
.15
.65
.15
.65
.65
.15
.65
.15
.65
.65
.15
.65
.15
.65
.65
.15
.65
.15
.65
.65
.15
.65
.15
.65
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
300.
300.
300.
300.
300.
300.
300.
300,
300.
300.
200.
200.
200.
200.
200.
220.
220.
220.
220.
220.
240.
240.
240.
240.
240.
260.
260.
260.
260.
260.
280.
280.
280.
280.
280.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
KM
ARC
ARC
ARC
ARC
ARC
ARC
ARC
ARC
ARC
ARC
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
COL
20.
30.
40.
50.
60.
70.
80,
90.
100.
110.
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
ROW
0
0
0
0
0
0
0
0
0
0
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
DEC
DEC
DEG
DEG
DEG
DEG
DEG
DBG
DEG
DEG
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
I. REPORT NO.
EPA-454/R-92-021
3, RECIPIENT'S ACCESSiON NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
October 1992
A Modeling Protocol for Applying MESOPUFF H to Long
Range Transport Problems
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gerald L. Gipson
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
t). CONTRACT/GRANT NO.
68-023733
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This guidance document describes recommended procedures for the application of MESOPUFF n to
long range transport problems, including a discussion of spatial and temporal scales of analysis,
compilation of meteorological data bases, application of MESOPUFF n preprocessors, and control
strategy evaluation. An example MESOPUFF n application for a single isolated power plant stack
located near the Ohio River in southeastern Ohio is presented. Example input data sets for two
meteorological preprocessor programs, READ56 and MESOPAC n, are presented in an appendix.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS^OPEN ENDED TERMS
c. COSAT] Field/Group
Air Pollution
Atmospheric Dispersion Modeling
Long Range Transport
Puff Models
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
63 (incl.
appendix)
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
El'A Form ZZ20-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |